Nature News reports that ResearchGate take-down notices are imminent and perhaps already occurring which threatens to lower the percent of papers that are available to freely view on the web. How much of an impact might this have?

I analyzed the recent Lazy Scholar data set described here for ResearchGate as a source of free full texts. Here is what I can tell:

  • When Lazy Scholar finds at least one full text, 17% of the time ResearchGate is one of the sources.
  • About 9% of the time, ResearchGate is the sole full text source and Lazy Scholar cannot find it elsewhere.

However this is likely a slight underestimation because I do not have resolved DOAI.io links in the records, which often have ResearchGate as the source. DOAI.io is a sole full text source only 1% of the time though, so let’s go with the high end and call it an even 10% where ResearchGate is a sole full text source.

Therefore, if my math is correct, assuming 40% of free texts have the potential to come down from ResearchGate based on the article, that may drop the effectiveness of Lazy Scholar from 36-41% to 34-39%. I’ll be monitoring Lazy Scholar records to see if this prediction is met in the short-term, and exploring ways to filter out papers that breach copyright.